Re: SATA ICH5 not detected at boot, mm-kernels

[Tejun Heo <htejun@gmail.com>]

Hello,

Andrew Morton wrote:
> On Sun, 5 Nov 2006 15:36:33 +0100
> Fabio Coatti <cova@ferrara.linux.it> wrote:
> 
>> Hi all; It seems that problems like this has been already reported, but not 
>> exactly the same, so maybe thsi can add some infos. Otherwise, sorry for the 
>> noise.
>>
>> Starting from 2.6.19-rc1-mm1 and up to rc4-mm2, at boot the kernel is unable 
>> to detect two sata disks, connected to a ICH5 controller. Latest mm working 
>> kernel seems to be 2.6.18-mm3; 2.6.19-rc4 works just fine.
>>
>> On rc-4 (vanilla) the log is this:
>> Nov  5 13:26:37 kefk libata version 2.00 loaded.
>> Nov  5 13:26:37 kefk ata_piix 0000:00:1f.2: version 2.00ac6
>> Nov  5 13:26:37 kefk ata_piix 0000:00:1f.2: MAP [ P0 P1 IDE IDE ]
>> Nov  5 13:26:37 kefk ACPI: PCI Interrupt 0000:00:1f.2[A] -> GSI 18 (level, 
>> low) -> IRQ 17
>> Nov  5 13:26:37 kefk ata: 0x170 IDE port busy
>> Nov  5 13:26:37 kefk ata: conflict with ide1
> 
> hm.  What does that mean?

It means that IDE layer claimed the port.  It can be overridden by 
combined_mode kernel parameter.

>> Nov  5 13:26:37 kefk PCI: Setting latency timer of device 0000:00:1f.2 to 64
>> Nov  5 13:26:37 kefk ata1: SATA max UDMA/133 cmd 0x1F0 ctl 0x3F6 bmdma 0xF000 
>> irq 14
>> Nov  5 13:26:37 kefk ata2: DUMMY
>> Nov  5 13:26:37 kefk scsi1 : ata_piix
>> Nov  5 13:26:37 kefk ata1.00: ATA-7, max UDMA/133, 488397168 sectors: LBA48 
>> NCQ (depth 0/32)
>> Nov  5 13:26:37 kefk ata1.00: ata1: dev 0 multi count 16
>> Nov  5 13:26:37 kefk ata1.01: ATA-7, max UDMA/133, 586114704 sectors: LBA48 
>> NCQ (depth 0/32)
>> Nov  5 13:26:37 kefk ata1.01: ata1: dev 1 multi count 16
>> Nov  5 13:26:37 kefk ata1.00: configured for UDMA/133
>> Nov  5 13:26:37 kefk ata1.01: configured for UDMA/133
>>
>>
>> With -mm kernels, I see only the third disk, but attached to a different 
>> controller (output from rc4 vanilla):
>>
>> Nov  5 13:26:37 kefk ACPI: PCI Interrupt 0000:03:03.0[A] -> GSI 19 (level, 
>> low) -> IRQ 18
>> Nov  5 13:26:37 kefk ata3: SATA max UDMA/100 cmd 0xF8804080 ctl 0xF880408A 
>> bmdma 0xF8804000 irq 18
>> Nov  5 13:26:37 kefk ata4: SATA max UDMA/100 cmd 0xF88040C0 ctl 0xF88040CA 
>> bmdma 0xF8804008 irq 18
>> Nov  5 13:26:37 kefk scsi3 : sata_sil
>> Nov  5 13:26:37 kefk ata3: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
>> Nov  5 13:26:37 kefk ata3.00: ATA-7, max UDMA/133, 625142448 sectors: LBA48 
>> NCQ (depth 0/32)
>> Nov  5 13:26:37 kefk ata3.00: ata3: dev 0 multi count 0
>> Nov  5 13:26:37 kefk ata3.00: configured for UDMA/100
>> Nov  5 13:26:37 kefk scsi4 : sata_sil
>> Nov  5 13:26:37 kefk ata4: SATA link down (SStatus 0 SControl 310)
>> Nov  5 13:26:37 kefk scsi 3:0:0:0: Direct-Access     ATA      Maxtor 6V320F0   
>> VA11 PQ: 0 ANSI: 5
>> Nov  5 13:26:37 kefk SCSI device sdc: 625142448 512-byte hdwr sectors (320073 
>> MB)
>> Nov  5 13:26:37 kefk sdc: Write Protect is off
>>
> 
> And why doesn't -mm report the same conflict?  I assume the .config is the
> same?

Also, please post full dmesg of both kernels.

Thanks.

-- 
tejun